Tokyo–Washington AI system crafts first zero-day exploit
PUBLISHED May 11, 2026, 4:51 PM ET
Read, Watch or Listen
Google’s Threat Analysis Group (TAG) has confirmed the first documented case of a zero‑day exploit fully designed and deployed by artificial intelligence, targeting a previously unknown vulnerability in the Linux kernel’s memory management system. The AI-driven attack framework, dubbed “Strix,” uses an automated toolset called “Hexstrike” to discover and weaponize software flaws without human coding. Activity was linked to advanced persistent threat group UNC2814 after a major East Asian cybersecurity firm detected anomalous traffic against Japanese and U.S. technology companies. Investigators say Strix located a TP‑Link OFTP firmware buffer overflow in under fifteen minutes, signaling a major shift in offensive cyber capabilities.
By James Porter | JQJO News
Timeline of Events
- Earlier this year, TAG begins coordinated investigation
- Soon after, East Asian firm flags anomalous traffic
- Subsequently, activity attributed to APT UNC2814
- During forensics, investigators identify Strix framework
- Shortly afterward, Hexstrike toolset behavior documented
- Same period, Linux kernel memory vulnerability uncovered
- Minutes later, TP-Link OFTP overflow exploit generated
- This week, Google publicly confirms AI-crafted zero-day
News Intelligence
- This AI-driven attack could change cybersecurity. It found and exploited a software flaw in minutes. That's faster than any human. Your devices could be at risk if they run on Linux. Check for updates regularly to stay safe.
Comments